KR100841591B1 - Apparatus and method for adaptive de-interlacing using takagi-sugeno fuzzy model - Google Patents

Abstract

An apparatus and a method for adaptive de-interlacing using a Takagi-Sugeno fuzzy model are provided to acquire the interpolation image of good image quality irrespective of the edge property or motion property of an image and apply an interpolation method selectively to the image in consideration of complexity according to the motion property of the image, thereby improving objective and subjective image quality and an algorithm calculating rate. A maximum difference coefficient calculator(110) calculates the first maximum difference coefficient of a difference value between the maximum value and the minimum value out of the pixel values of first pixels selected in a special axis window set to a current frame including a interpolation target pixel, and calculates the second maximum difference coefficient of a difference value between the maximum value and the minimum value out of the pixel values of second pixels selected in a temporal axis window set to reference frames near on the temporal axis for the current frame including the interpolation target pixel. A membership value calculator(120) calculates a first membership value for each of edge direction angles which the first pixels form in the spatial axis window on the basis of the first maximum difference coefficient, and calculates a second membership value for each of edge direction angles which the second pixels form in the temporal axis window on the basis of the second maximum difference coefficient. An edge direction angle determining unit(130) determines an edge direction angle corresponding to the maximum membership value out of the first membership value and the second membership value as the maximum edge direction angle. A first interpolation unit(170) interpolates the interpolation target pixel by the average pixel value of pixels positioned on the maximum edge direction angle in the window to which pixels corresponding to the maximum membership value belong.

Description

Apparatus and method for adaptive de-interlacing using Takagi-Sugeno fuzzy model}

1 is a block diagram showing a detailed configuration of a preferred embodiment of an adaptive deinterlacing apparatus according to the present invention;

2 is a diagram illustrating a space axis and a time axis window set for a current frame including an interpolation target pixel and a frame temporally adjacent to the current frame;

3 is a diagram illustrating a membership function for obtaining a first membership value and a second membership value;

FIG. 4 is a diagram illustrating the belonging values of a fuzzy set and each fuzzy set according to a Takagi-Sugeno (TS) fuzzy model. FIG.

5 is a view showing the implementation of a preferred embodiment for the adaptive de-interlacing method according to the present invention, and

6A to 6D are images obtained by a method of adaptively interpolating according to a layer to which a Bob method, a Weave method, an FDIO method, and a fuzzy variable belong to an experiment image (Takagi-Sugino Interpolation: TSI), respectively. Is shown.

The present invention relates to an adaptive deinterlacing apparatus and method, and more particularly, to an adaptive deinterlacing apparatus and method using the Takagi-Sugeno fuzzy model.

Most video sources (DVD, SDTV, HDTV) are transmitted by interlaced scanning, while receiver displays such as DLP, LCD, LCOS (SXRX, D-ILA) and PDP are progressive scans. Since an image is reproduced by a progressive scanning method, a de-interlacing technology has been developed to solve this problem.

De-interlacing technology is a scanning format conversion technology that converts an image input by a parallel scan into a progressive scan image and outputs it. A representative example in which the deinterlacing technique is used is a case where a television signal, which is an air wave broadcast, is viewed on a computer. This deinterlacing technique predicts pixel values missing in a field to make a complete frame. De-interlacing techniques can be classified into two types, inter-field interpolation and intra-field interpolation.

One example of interfield interpolation is Weave, which combines top-field and bottom-field to form a frame. That is, one frame is implemented by simply inserting the line of the immediately preceding field between the lines of the current field. This can be easily implemented when interpolating an image without motion compensation, but when interpolating an image with motion compensation, horizontal lines are generated or the screen is degraded. In-field interpolation is also commonly known as Bob, which creates a frame using each line of a field twice. In other words, a new frame is implemented by inserting average data of two lines in an area between two lines in one field. This intrafield interpolation method prevents horizontal lines from appearing in the image with motion compensation. However, when the still image is interpolated, the screen is deteriorated, and the complex and fine screen is shaken at 30 Hz.

Currently, the edge-based interpolation (ELA) interpolation method using the average of pixel values located on the upper and lower lines of the pixel to be interpolated and the direction-oriented interpolation (DOI) interpolation method considering the direction of the boundary part are most widely used. Such interpolation method still needs improvement of interpolated image quality and calculation speed.

The technical problem to be achieved by the present invention is to obtain an interpolation image with good image quality regardless of the edge characteristics or motion characteristics of the image, and at the same time, to selectively apply the interpolation method in consideration of the complexity according to the motion characteristics of the image. The present invention provides an apparatus and method for adaptive deinterlacing that can improve image quality and algorithm calculation speed.

Another technical problem to be solved by the present invention is to obtain an interpolation image having good image quality regardless of edge characteristics or motion characteristics of an image, and to simultaneously apply an interpolation method selectively considering the complexity according to the motion characteristics of the image. The present invention provides a computer-readable recording medium for executing a computer on an adaptive deinterlacing method which can improve subjective picture quality and algorithm calculation speed.

In order to achieve the above technical problem, a preferred embodiment of the adaptive deinterlacing apparatus according to the present invention is the maximum among pixel values of a plurality of first pixels selected within a spatial axis window set for a current frame including an interpolation target pixel. A first maximum coefficient, which is a difference between a value and a minimum value, is calculated, and is the maximum among pixel values of a plurality of second pixels selected within a time axis window set for reference frames adjacent to the current frame including the interpolation target pixel and a time axis; A maximum difference coefficient calculating unit for calculating a second maximum difference coefficient that is a difference between the value and the minimum value; A first membership value for each edge direction angle formed by the first pixels in the space axis window is calculated based on the first maximum difference coefficient, and second pixels in the time axis window are calculated based on the second maximum difference coefficient. A membership value calculator configured to calculate a second membership value for each edge direction angle formed; An edge direction angle determiner that determines an edge direction angle corresponding to a maximum membership value among the first membership values and the second membership values as a maximum edge direction angle; And a first interpolation unit which interpolates the interpolation target pixel by an average pixel value of pixels positioned on the maximum edge direction angle in a window to which pixels corresponding to the maximum membership value belong.

In order to achieve the above technical problem, a preferred embodiment of the adaptive deinterlacing method according to the present invention includes a pixel value of a plurality of first pixels selected within a spatial axis window set for a current frame including an interpolation target pixel. Calculating a first maximum difference coefficient that is a difference between a maximum value and a minimum value; Calculating a second maximum difference coefficient that is a difference between a maximum value and a minimum value among pixel values of a plurality of second pixels selected within a time axis window set for the current frame including the interpolation target pixel and adjacent reference frames on a time axis; ; A first membership value for each edge direction angle formed by the first pixels in the space axis window is calculated based on the first maximum difference coefficient, and second pixels in the time axis window are calculated based on the second maximum difference coefficient. Calculating a second membership value for each edge direction angle; Determining an edge direction angle corresponding to a maximum membership value among the first membership values and the second membership values as a maximum edge direction angle; And a first interpolation step of interpolating the interpolation target pixel by an average pixel value of pixels positioned on the maximum edge direction angle in a window to which pixels corresponding to the maximum membership value belong.

By applying the Takagi-Sugeno Fuzzy Theory when converting a progressive scan type video signal into a progressive scan type video signal, an area with a lot of edges and a small amount of motion at the same time is present. For an image including all regions, an image having an objectively and subjectively improved image quality can be obtained than an image reconstructed by a conventional Weave or Bob method.

DETAILED DESCRIPTION Hereinafter, exemplary embodiments of an adaptive deinterlacing apparatus and method according to the present invention will be described in detail with reference to the accompanying drawings.

1 is a block diagram showing a detailed configuration of a preferred embodiment of the adaptive deinterlacing apparatus according to the present invention.

Referring to FIG. 1, the adaptive deinterlacing apparatus according to the present invention includes a maximum difference coefficient calculating unit 110, a membership value calculating unit 120, an edge direction angle determining unit 130, a fuzzy set classification unit 140, A fuzzy variable calculating unit 150, an interpolation means selecting unit 160, a first interpolation unit 170, a second interpolation unit 180, and a third interpolation unit 190 are provided.

The maximum difference coefficient calculating unit 110 calculates a first maximum difference coefficient that is a difference between a maximum value and a minimum value among pixel values of a plurality of first pixels selected within a spatial axis window set for a current frame including an interpolation target pixel. The second maximum difference coefficient, which is a difference between a maximum value and a minimum value, is calculated among pixel values of a plurality of second pixels selected in a time axis window set for the current frame including the interpolation target pixel and adjacent reference frames on the time axis. . 2 shows a space axis and a time axis window set for a current frame including an interpolation target pixel and a frame temporally adjacent to the current frame.

As shown in FIG. 2, the maximum difference coefficient calculating unit 110 performs an interpolation on the current frame 210 including the interpolation target pixel in order to interpolate the lost pixel value x (i, j, k) of the interpolation target pixel. Based on the six selected pixels 212, 214, 216, 218, 220, and 222 in the set space axis window 250, a first maximum difference coefficient is calculated by the following equation.

는 공간축 윈도우 내에 위치하는 화소의 개수이다.In Equation 1, SMDW is a first maximum difference coefficient, which is the maximum value of the difference between pixel values between interpolation target pixels located in the space axis window 250 and adjacent pixels, and β is a value of SMDW from 0 to 255. Proportional constant to limit,

Is the number of pixels located in the space axis window.

In addition, the maximum difference coefficient calculating unit 110 includes six pixels 232, selected in the time axis window 250 set for the front and rear frames 220 and 240 adjacent to the current frame 210 including the interpolation target pixel. Based on 234, 236, 242, 244, and 246, the second maximum difference coefficient is calculated by the following equation.

는 시간축 윈도우 내에 위치하는 화소의 개수이다.In Equation 2, TMDW is a second maximum difference coefficient which is the maximum value of the difference between pixel values between the interpolation target pixel and the adjacent pixels located in the time axis window 260, and β is a limit value of TMDW from 0 to 255. Is proportional to

Is the number of pixels located in the time axis window.

The membership value calculation unit 120 uses the following equation for each slope of θ (0 °, ± 45 °, ± 60 °) to apply the fuzzy inference method based on the first maximum difference coefficient. A first membership value for each edge direction angle formed by the first pixels in 250 is calculated.

In addition, the membership value calculator 120 uses the following equation for each slope of θ (0 °, ± 45 °, ± 60 °) to apply the fuzzy inference method based on the second maximum difference coefficient. A second membership value for each edge direction angle formed by the second pixels in 260 is calculated.

및

는 각각 제1멤버쉽값 및 제2멤버쉽값이고,

및

는 각각

와

의 θ가 0°일 때 공간축 윈도우 및 시간축 윈도우에 각각 대응하는 제1최대차계수 및 제2최대차계수이고,

및

는 각각 공간축 윈도우와 시간축 윈도우에서의 기울기 θ와 수평 위치차 d(-1 또는 0 또는 1)에 대한 보간할 화소에 인접하여 위치하는 화소들의 차이값이다.In Equations 3 and 4,

And

Are the first membership value and the second membership value, respectively,

And

Are each

Wow

Is a first maximum coefficient and a second maximum coefficient corresponding to the space axis window and the time axis window, respectively, when θ is 0 °,

And

Is a difference value between pixels positioned adjacent to the pixel to be interpolated for the tilt θ and the horizontal position difference d (−1 or 0 or 1) in the space axis window and the time axis window, respectively.

,

,

및

는 도 3에 도시된 멤버쉽 함수로부터 정 의된다. 표 1에는 각각의 멤버쉽 값을 구하기 위한 파라미터가 기재되어 있다.The membership function for obtaining the first membership value and the second membership value is a function corresponding to a fuzzy set small as shown in FIG. 3, and a parameter for obtaining each membership value.

,

,

And

Is defined from the membership function shown in FIG. Table 1 lists the parameters for obtaining each membership value.

Reference parameter value A b _{0 °} b ₀ B b _{± 45 °} 3/4 × b ₀ C b _{± 60 °} 1/2 × b ₀ D a _{0 °} 1/4 × b ₀ E a _{± 45 °, ± 60 °} 1/5 × b ₀

및

는 다음의 수학식으로 표현된다.The edge direction angle determiner 130 determines the edge direction angle corresponding to the maximum membership value among the first and second membership values as the maximum edge direction angle. At this time, the input value input to the edge direction angle determiner 130

And

Is expressed by the following equation.

및

는 각각 공간축 윈도우와 시간축 윈도우에서 θ에 의해 결정되는 경사 차이값으로 0, ±1, ±2의 값을 갖는다.In Equations 5 and 6, d is a horizontal component difference value of 0, 1, -1,

And

Are the slope difference values determined by θ in the space and time axis windows, respectively, and have values of 0, ± 1, and ± 2.

The edge direction angle determiner 130 fuzzy the input values represented by Equations 5 and 6 to find the membership value and the fuzzy sets. The edge direction angle determiner 130 obtains the minimum value of the fuzzy set SMALL obtained from the input value and the membership function for each value of θ through an inference engine, and the following equation. This process is performed for the space axis and the time axis respectively.

In equations (7) and (8), C _m and D _m are the minimum values obtained for the space axis and the time axis, respectively, and A _m (·) and B _m (·) are the values in which the difference values on the space axis and the time axis are respectively purged. to be.

Next, the edge direction detection unit 130 compares all the membership values in the five directions obtained by the equations (7) and (8) and calculates the θ value corresponding to C _m (·) or D _m (·) having the maximum membership value. Find. Through the above process, the edge direction detection unit 130 obtains the maximum membership value expressed by the following equation, and determines the θ value at that time in the edge direction.

The first interpolator 170 interpolates the interpolation target pixel by an average pixel value of pixels positioned on the maximum edge direction angle in a window to which pixels corresponding to the maximum membership value belong. Interpolation by the first interpolation unit 170 is performed by the following equation, and this interpolation method may be referred to as FDOI (Fuzzy Directional-oriented Interpolation) interpolation.

As described above, the FDOI interpolation method has an objectively and subjectively improved image quality compared to the image reconstructed by the conventional Weave method or Bob method for an image including both an area with high edges and an area with low motions while edges exist. There is an advantage to get it.

The interpolation method that is suitable among the Weave method, the Bob method, and the FDOI method for each area in consideration of the complexity according to the motion characteristics of each area of the current frame including the interpolation target pixel using the Takagi-Sugeno fuzzy law is described below. This section describes the process of performing interpolation by determining.

The fuzzy set classification unit 140 classifies the first maximum coefficient and the second maximum differential coefficient into SMALL and LARGE fuzzy sets based on a predetermined reference value (for example, 2 ⁴ ). . 4 shows the belonging values of the fuzzy set and each fuzzy set according to the Takagi-Sugeno (TS) fuzzy model.

The fuzzy variable calculation unit 150 converts the input values SMDW and TMDW into fuzzy variables by using the TS fuzzy model representing the overall fuzzy model of the system as the fuzzy combination of the respective linear system models.

Generalizing the equation (11), the TS fuzzy model can be expressed as the following equation.

는 퍼지변수, SMDW는 제1최대차계수, TMDW는 제2최대차계수, a_ij는 TS 계수 행렬, h_i[z(t)]는 정규화된 가중치값으로

이고,

이며, z_j(t)는 제1최대차계수 또는 제2최대차계수, 그리고, A_ji[z_j(t)]는 제1최대차계수 또는 제2최대차계수가 속하는 퍼지셋의 소속값이다. In Equation 12,

Is the fuzzy variable, SMDW is the first maximum coefficient, TMDW is the second maximum coefficient, a _ij is the TS coefficient matrix, and h _i [z (t)] is the normalized weight value.

ego,

Where z _j (t) is the first or second maximum coefficient, and A _ji [z _j (t)] is the belonging value of the fuzzy set to which the first or second maximum coefficient belongs. to be.

In this case, the TS coefficient matrix a _ij of Equation 12 may be defined as follows.

The interpolation means selecting unit 160 divides the fuzzy variables calculated by the fuzzy variable calculating unit 150 into three hierarchies listed in Table 2 based on the magnitude, and interpolates by the equation (14) according to the hierarchical layer to which the fuzzy variable belongs. Optionally determine the interpolation means of the target pixel.

hierarchy division Tier 0 low TMDW (value ≤ t ₁ ) Tier 1 large SMDW (t ₁ <membership value ≤ t ₂ ) Tier 2 high TMDW & SMDW (t ₂ <membership value)

In this case, t ₁ and t ₂ are determined experimentally, and may be set to 37 and 60, for example.

According to Equation 14, when the layer to which the interpolation target pixel belongs is 0, the interpolation is performed by the Weave method, the 1 by the Bob method, and the 2 by the FDOI method. For this selective interpolation, the second interpolator 180 performs interpolation in a Weave method of interpolating the interpolation target pixel by the pixels selected in the time axis window, and the third interpolator 190 operates in the space axis window. The interpolation is performed in a Bob manner in which interpolation target pixels are interpolated by the selected pixels.

FIG. 5 is a flowchart illustrating a preferred embodiment of the adaptive deinterlacing method according to the present invention.

Referring to FIG. 5, the maximum difference coefficient calculating unit 110 is a difference value between a maximum value and a minimum value among pixel values of a plurality of first pixels selected within a spatial axis window set for a current frame including an interpolation target pixel. One maximum difference coefficient is calculated (S510). In addition, the maximum coefficient calculation unit 110 is a difference between the maximum value and the minimum value among the pixel values of the plurality of second pixels selected within the time axis window set for the current frame including the interpolation target pixel and the reference frames adjacent to each other on the time axis. The second maximum difference coefficient is calculated (S520). Next, the membership value calculator 120 calculates a first membership value for each edge direction angle formed by the first pixels in the space axis window based on the first maximum difference coefficient, and based on the second maximum difference coefficient. A second membership value for each edge direction angle formed by the second pixels in the time axis window is calculated (S530). The edge direction angle determiner 130 determines the edge direction angle corresponding to the maximum membership value among the first membership values and the second membership values as the maximum edge direction angle (S540).

Next, the fuzzy set classification unit 140 classifies the first maximum coefficient and the second maximum coefficient into small purge sets and large purge sets based on a predetermined reference value (for example, 2 ⁴ ) (S550). . Subsequently, the fuzzy variable calculation unit 150 converts the first and second maximum coefficients into fuzzy variables according to Equation 12 according to the fuzzy set to which the first and second maximum coefficients belong (S560). ). In addition, the interpolation means selecting unit 160 divides the fuzzy variables calculated by the fuzzy variable calculating unit 150 into three hierarchies listed in Table 2 based on the magnitude, and according to Equation 14 according to the hierarchical layer to which the fuzzy variables belong. An interpolation method of the interpolation target pixel is selectively determined among the weave method, the Bob method, and the FDOI method (S570). Finally, the interpolation unit determined by the interpolation means selection unit 160 among the first interpolation unit 170, the second interpolation unit 180, and the third interpolation unit 190 interpolates the interpolation target pixel (S580).

6A to 6D show images obtained by a method of adaptively interpolating according to a layer to which a Bob method, a Weave method, an FDIO method, and a fuzzy variable belong to an experimental image (Takagi-Sugino Interpolation: TSI), respectively. Is shown.

6A to 6D, the image interpolated by the Bob method is blurred as a whole, and the image interpolated by the Weave method has a serrated distortion in too many parts in a high-motion region. appear. In contrast, the image interpolated by the FDOI method has better overall image quality than the image interpolated by the Bob and Weave methods. In addition, the interpolation method that is performed by adaptively selecting interpolation method using Takagi-Sugeno fuzzy model is performed by Bob method for low edge components, Weave method for low motion components, and areas with many edge components or By performing interpolation by the FDOI method, an area with a lot of motion can obtain an image having good image quality objectively and subjectively. Table 3 lists the PSNR (in dB) for each interpolation method for different experimental images.

Interpolation method Experimental video Akiyo Flower Mobile News T.Tennis ELA 37.9311 21.6810 23.5328 31.4749 27.4082 Bob 39.8581 22.1905 25.5118 33.6152 28.5659 DOI 39.7692 22.0041 24.8518 33.3184 27.8734 FDED 40.0392 22.0684 25.2252 33.5481 28.4051 Weave 43.7858 20.2949 23.5373 36.4714 27.9963 VTMF 42.3840 22.3817 25.2296 37.3808 29.5599 STELA 44.6554 22.9906 27.2602 39.2848 31.5877 EDT 44.1538 21.8967 25.1201 36.9747 28.9696 FDOI 47.8423 23.2425 28.2957 40.7563 31.8844 TSI 47.7379 23.1615 28.1912 40.6664 31.7846

Referring to Table 3, it can be seen that the FDOI method and the TSI method perform better for all images than other interpolation methods. In addition, in the above-described embodiment, the interpolation target pixel is interpolated by the interpolation method selected according to the characteristics of the interpolation target pixel among the Weave, Bob and FDOI interpolation methods. The interpolation method of may be selected.

 The invention can also be embodied as computer readable code on a computer readable recording medium. The computer-readable recording medium includes all kinds of recording devices in which data that can be read by a computer system is stored. Examples of computer-readable recording media include ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical data storage, and the like, and may also be implemented in the form of a carrier wave (for example, transmission over the Internet). Include. The computer readable recording medium can also be distributed over network coupled computer systems so that the computer readable code is stored and executed in a distributed fashion.

Although the preferred embodiments of the present invention have been shown and described above, the present invention is not limited to the specific preferred embodiments described above, and the present invention belongs to the present invention without departing from the gist of the present invention as claimed in the claims. Various modifications can be made by those skilled in the art, and such changes are within the scope of the claims.

According to the adaptive de-interlacing apparatus and method according to the present invention, when the image signal of the progressive scan method is converted into the image signal of the forward scan method, the edge is applied by applying Takagi-Sugeno Fuzzy Theory. It is possible to obtain an image having an objectively and subjectively improved image quality compared to the image reconstructed by the conventional Weave method or the Bob method, for an image including both a high motion region and a low motion region at the same time. In addition, by applying the Takagi-Sugeno Fuzzy Theory, the pixels to be interpolated are divided into three layers by using the maximum differences between the pixels in the window, which are the time and space axes around the pixels to be interpolated. The interpolation method can be selectively applied in consideration of the complexity according to the characteristics of the motion of the layer to improve the objective and subjective picture quality and the algorithm calculation speed.

Claims (11)

A first maximum difference coefficient, which is a difference between a maximum value and a minimum value among pixel values of a plurality of first pixels selected within a spatial axis window set for the current frame including the interpolation target pixel, is calculated and includes the interpolation target pixel. A maximum difference coefficient calculating unit configured to calculate a second maximum difference coefficient that is a difference between a maximum value and a minimum value among pixel values of a plurality of second pixels selected within a time axis window set for reference frames adjacent to a current frame and a time axis; A first membership value for each edge direction angle formed by the first pixels in the space axis window is calculated based on the first maximum difference coefficient, and second pixels in the time axis window are calculated based on the second maximum difference coefficient. A membership value calculator configured to calculate a second membership value for each edge direction angle formed; An edge direction angle determiner that determines an edge direction angle corresponding to a maximum membership value among the first membership values and the second membership values as a maximum edge direction angle; And And a first interpolator configured to interpolate the interpolation target pixel by an average pixel value of pixels positioned on the maximum edge direction angle in a window to which pixels corresponding to the maximum membership value belong. Deinterlacing device. The method of claim 1, A fuzzy set classification unit for classifying the first maximum coefficient and the second maximum coefficient into small purge sets and large purge sets based on reference values set or experimentally determined by a user's experience; A fuzzy variable calculator for converting the first maximum coefficient and the second maximum coefficient into fuzzy variables according to a fuzzy set to which the first maximum coefficient and the second maximum coefficient belong; And And an interpolation means selector for selectively determining an interpolation means of the interpolation target pixel based on the magnitude of the fuzzy variable.

, here,

Is the fuzzy variable, SMDW is the first maximum coefficient, TMDW is the second maximum coefficient, a _ij is the Takagi-Sugeno coefficient matrix, and h _i [z (t)] is a normalized weight value.

ego,

Where z _j (t) is the first maximum coefficient or the second maximum coefficient, and A _ji [z _j (t)] is the fuzzy to which the first maximum coefficient or the second maximum coefficient belongs. The membership of the set. The method of claim 2, A second interpolation unit which interpolates the interpolation target pixel by pixels selected in the time axis window; And And a third interpolation unit which interpolates the interpolation target pixel by the pixels selected in the space axis window. The interpolation means selecting unit selects the second interpolation unit as interpolation means of the interpolation target pixel when the magnitude of the fuzzy variable is equal to or less than an experimentally determined first threshold value, and the magnitude of the fuzzy variable is larger than the first threshold value. The third interpolation part is selected as an interpolation means of the interpolation target pixel if the second threshold value is larger than or equal to an experimental value. If the magnitude of the fuzzy variable is larger than the second threshold value, the first interpolation part is the interpolation object. Adaptive de-interlacing apparatus characterized in that the selection by the interpolation means of the pixel. The method according to claim 1 or 2, The membership value calculating unit calculates the first membership value and the second membership value according to the following equation:

,

, here,

And

Are each of the first membership value and the second membership value,

,

,

And

Is a parameter for determining each membership value obtained from the membership function set for the small fuzzy set and the large fuzzy set.

Wow

When the θ of 0 is 0 °, the first maximum coefficient and the second maximum coefficient corresponding to the space axis window and the time axis window are respectively

And

ego,

And

Are difference values of pixels positioned adjacent to the pixel to be interpolated with respect to the tilt θ and a horizontal position difference d (−1 or 0 or 1) in the space axis window and the time axis window, respectively. The method according to claim 1 or 2, And the maximum difference coefficient calculating unit calculates the first maximum difference coefficient and the second maximum difference coefficient according to the following equation:

,

, Here, SMDW and TMDW are the first and second maximum coefficients, respectively, β is a proportional constant for limiting the SMDW and TMDW value from 0 to 255,

And

Are the number of pixels located in the space axis window and the time axis window, respectively, and x (i, j, k) are lost pixels located at (x, y) in the current frame corresponding to the kth frame. Calculating a first maximum difference coefficient that is a difference between a maximum value and a minimum value among pixel values of a plurality of first pixels selected within a spatial axis window set for a current frame including an interpolation target pixel; Calculating a second maximum difference coefficient that is a difference between a maximum value and a minimum value among pixel values of a plurality of second pixels selected within a time axis window set for the current frame including the interpolation target pixel and adjacent reference frames on a time axis; ; A first membership value for each edge direction angle formed by the first pixels in the space axis window is calculated based on the first maximum difference coefficient, and second pixels in the time axis window are calculated based on the second maximum difference coefficient. Calculating a second membership value for each edge direction angle; Determining an edge direction angle corresponding to a maximum membership value among the first membership values and the second membership values as a maximum edge direction angle; And And a first interpolation step of interpolating the interpolation target pixel by an average pixel value of pixels positioned on the maximum edge direction angle in a window to which pixels corresponding to the maximum membership value belong. Enemy deinterlacing method. The method of claim 6, Classifying the first maximum coefficient and the second maximum coefficient into small purge sets and large purge sets based on reference values set or experimentally determined by a user's experience; Converting the first maximum coefficient and the second maximum differential coefficient into fuzzy variables according to the following equation according to the fuzzy set to which the first maximum coefficient and the second maximum coefficient belong; And And selectively determining an interpolation method of the interpolation target pixel based on the magnitude of the fuzzy variable. The adaptive deinterlacing method further comprises:

, here,

Is the fuzzy variable, SMDW is the first maximum coefficient, TMDW is the second maximum coefficient, a _ij is the Takagi-Sugeno coefficient matrix, and h _i [z (t)] is a normalized weight value.

ego,

Where z _j (t) is the first maximum coefficient or the second maximum coefficient, and A _ji [z _j (t)] is the fuzzy to which the first maximum coefficient or the second maximum coefficient belongs. The membership of the set. The method of claim 7, wherein A second interpolation step of interpolating the interpolation target pixel by pixels selected in the time axis window; And And a third interpolation step of interpolating the interpolation target pixel by the pixels selected in the space axis window. In the selecting of the interpolation means, if the magnitude of the fuzzy variable is less than or equal to an experimentally determined first threshold value, performing the second interpolation step to interpolate the interpolation target pixel, and the magnitude of the fuzzy variable is greater than the first threshold value. The interpolation target pixel is interpolated by performing the third interpolation step when the second threshold value is larger than or equal to the experimentally determined value. When the size of the fuzzy variable is larger than the second threshold value, the interpolation is performed by performing the first interpolation step. Adaptive deinterlacing method characterized by interpolating a target pixel. The method according to claim 6 or 7, Adaptive deinterlacing method, characterized in that for calculating the membership value in the step of calculating the first membership value and the second membership value by the following equation:

,

, here,

And

Are each of the first membership value and the second membership value,

,

,

And

Is a parameter for determining each membership value obtained from the membership function set for the small fuzzy set and the large fuzzy set.

Wow

When the θ of 0 is 0 °, the first maximum coefficient and the second maximum coefficient corresponding to the space axis window and the time axis window are respectively

And

ego,

And

Are difference values of pixels positioned adjacent to the pixel to be interpolated with respect to the tilt θ and a horizontal position difference d (−1 or 0 or 1) in the space axis window and the time axis window, respectively. The method according to claim 6 or 7, Adaptive deinterlacing method characterized in that for calculating the maximum difference coefficient in the step of calculating the first maximum coefficient and the second maximum difference coefficient by the following equation:

,

, Here, SMDW and TMDW are the first and second maximum coefficients, respectively, β is a proportional constant for limiting the SMDW and TMDW value from 0 to 255,

And

Are the number of pixels located in the space axis window and the time axis window, respectively, and x (i, j, k) are lost pixels located at (x, y) in the current frame corresponding to the kth frame. A computer-readable recording medium having recorded thereon a program for executing the adaptive deinterlacing method according to claim 6 or 7.

Images